Ok, since thievery is here to stay, there shouldn't be any reason that it can't be productive or acceptable. If we just try and understand the principles and motivations for it, any modifications in the future could be better. So first, a why and how stealing peoples junk.
Leverage: This is a distinction between thieves and griefers. Thieves will generally steal from people who have more than they do. This is a form of competition, the further that players progress, the more they thieves will want to rob them. This can be productive for the game world as a dynamic environment; on a long enough timeline, certain playing styles will outpace the majority of other players and disporportional tactics like thievery can diminish the less productive byproducts of an otherwise unchecked top percentile.
The Path of Least Resistance: A thief will prefer to go through the front door if they needn't smash your wall down. They would steal your money before your resources if resources weren't the only object of trade right now. The problems with thievery right now come from the fact the most successful way to steal from an advanced player is very close to outright griefing. Thieves can't pick a lock one time, they have to break the wall down steal your key and render it useless for much longer. They can't steal just the products of hard work, but the process of work itself.
To resolve issues of organized vandalism combined with burglary, we should allow thieves to steal in a proper way. There needs to be specialized counters to defenses, lockpicking gates, climbing, or pickpocketing keys. It shouldn't be common for thieves to employ tactics typically used by invading armies for assaulting the front of a castle; if we were to continue this tactic it would appropriate for players to place pots of boiling oil over their walls or dig moats.
Secondly, there needs to be a reasonable defense of ranging. The defense of throwing away characters is not a good one and is hardly any consolation to those who have been burgled. Scents need to be less permanent and indestructable and have a reasonable probability of leading to a beneficial conclusion (something better than symbolic vengeance kill). You're never going to catch your actual thief if they know you're coming and there's nothing they can do about it. It might be better to have scents that also lead to the objcts that were stolen and introduce a system to fence them so that thieves can get away with it if they know what they're doing. Then it would be better to remove the murder scents and summoning altogether, if you can track your thief and your objects you have a greater chance to get something out of your efforts. Thieves could defend themselves by altering scents and fencing goods; I mentioned pickpocketing earlier, it might work best if a thief can steal an object that has another player's scent and use it in some way to mask their own. For example if I were to steal some wool from a player's inventory, make a toga and wear it as I rob their house, it might cover up or diminish the theft scent I would leave; likewise if I stole from a random stranger, they could possibly be framed if I used their stuff to mask my scent. It could leave a false scent, two scents, or no scent depending whatever variables seem appropriate. However, objects should always leave scent when stolen, up until they are successfully fenced, which I imagine should be moderately costly (to deter small theft against players who have little to begin with).
Also, there needs to be a currency. I know we have coin presses, that's good, but an actual monetary system will protect people's work and allow thieves to carry on with their business. If there is a currency, that is, a standard monetary system of exchange for goods and services, thieves will find acquiring money to be a more efficient goal for their trade (they can steal more with one person, faster and easier). It takes a good while to train 20 thieves to trespass and steal, and they have a limited lifespan. The problem now is that no one has money, and the only purpose that there is for coins is to add more labor to a valuable good. People would use currency if they had access to it, could buy things with it, and could carry it easily.
We can kill two birds with one stone by facilitating a salvage industry, an exchange for salvaged materials that can be resold. The world accumulates so much junk and abandonement and simply destroying it is a waste of the original materials. If we introduce a skill that salvage 1/2 of the original materials of baskets, frames, jars, houses, tables, etc. and then allow players to sell them to an exchange that pays in coins, a currency actually stands a chance of being adopted. Especially if the exchange can resell the raw materials for currency. This would clean up a lot of junk, soften the resource waste, and establish a currency with an observable value. The problem I see is just determining a fair market value for exchanged goods, it seems to fluctuate to some degree right now, and metals have a highly inflated value (which can be traced to the fact that they can't be reused or salvaged at all right now).
I think altogether something like this can go a long way to resolve the current issues regarding thievery. The most important objective is to make it less attractive to use many disposable alts than it is to use one alt. I think if thieves can expect to get away with something on their own merit, they might do so; this would be better than, say, locking accounts to IPs which would just countered with proxies and unsecured access points. The goal should be to keep thievery in the game and not in the client design. In brief, an accepted currency that can buy resources and objects would give thieves a better target than resources and objects themselves, an exchange would balance the market and make thievery less destructive, and specifically designed counter-defenses would curb vandalism so long as they are easier to learn than strength is to train.